Genetic alterations predisposing to ovarian cancer, both among women in high-risk families and among patients in the general population, can be identified by positional cloning; integrating a systematic survey of genetic alterations in tumors, linkage mapping in families, physical mapping of the linked region, and comparison of sequences of candidate genes therein in tumors, patients and controls. We have successfully applied this approach to breast cancer and propose to do the same for ovarian cancer. Families with multiple cases of ovarian cancer will be identified from four cohorts and first tested for linkage to BRCA1. For families with ovarian cancer linked to 17q21, we will determine whether ovarian cancer is associated with different mutations of BRCA1 than is breast cancer; these will be compared with somatic alterations in sporadic ovarian tumors. We will evaluate epidemiologic risk factors that might distinguish ovarian cancer families with disease linked vs not linked to BRCA1. For families with ovarian cancer not linked to 17q21, we will map the critical genes by a genome-wide search, initially focusing on chromosomal regions with imbalance in ovarian tumors, and those with genes linked to predisposition to other forms of cancer also seen in the ovarian cancer families. We will identify candidate regions both from published reports of loss of heterozygosity (LOH), gene amplification, increased modes of gene expression and chromosomal alterations in ovarian tumors, and by screening malignant and benign ovarian tumors for LOH on all chromosome arms and karyotyping incident cases. The patterns of acquired mutations leading to ovarian cancer in the general population will be evaluated by identifying mutations at predisposing genes, LOH at critical regions and increased gene amplification and expression of ovarian oncogenes in the tumors of 300 previously identified patients, and 120 patients to be identified over the next four years. Correlations between genetic markers and tumor progression and outcome will also be made. Ovarian cancer is one of the most lethal cancers for women because it has been nearly impossible to detect at early, curable stages. Identifying genes critical to ovarian cancer development and progression, both in high risk families and among women generally, will pave the way for possible prevention, much earlier diagnosis and improved outcome.